Finance · NLP

How AI Improved Deal Sourcing for a Private Equity Fund

An end-to-end NLP solution that lets private equity funds and venture capital firms find and cluster relevant companies by meaning rather than by industry code or keyword.

1 week of analyst work saved per new niche market
Illustration of AI-assisted deal sourcing
Client
Private Equity Fund
Industry
Finance
Technology
NLP
Also applicable in
Venture Capital
01

The challenge

Deal sourcing helps investors find niche markets and assess the financial state of the companies in them, but the process is very time-consuming. It is made harder by the limits of the industry classification system used in the European Union.

Every company registered in the German Commercial Registry must describe its main business activities, yet EU classification systems such as NACE (Nomenclature of Economic Activities) do not cover new and emerging fields. There is, for example, no relevant code for companies focused on machine learning and artificial intelligence. Keyword search does not help much either, because it does not return companies that are semantically relevant to the query. As a result, private equity funds and business development companies struggle to explore niche markets and find the companies within them.

02

Our solution

We developed an end-to-end solution based on Natural Language Processing techniques and deep learning models. It collects, processes, analyses and displays data from different sources, and lets users run semantic and syntactic searches, explore clusters of companies and review similar companies.

A scraper extracts information from public data sources into a database holding each company’s name, registration date, status (active or inactive) and description of commercial activity. Pre-trained deep learning models such as multilingual BERT vectorise the company descriptions, so that companies can be clustered by semantic similarity. Semantic search works the same way: the search term is vectorised and matched against the closest descriptions in the database.

  • Scraper that aggregates public registry data into a company database
  • Vectorisation of company descriptions with multilingual BERT
  • Clustering of companies by semantic similarity, with user control over cluster size and similarity threshold
  • Semantic search that returns the companies closest in meaning to a search term
  • Extraction and aggregation of financial data to track companies, individually or as a group, over time
  • Export of the results as a PDF report
Tool dashboard: clusters of companies grouped by semantic similarity on the left, and the companies in the selected cluster with relevance scores on the right
03

The outcome

1 week of analyst work saved per new niche market

The client, a private equity company, gained an advanced toolset to explore the market and get insights into specific industries. The solution made deal sourcing more efficient, reducing an analyst’s workload by a week for every new niche market discovered.

The same approach is useful beyond private equity: it can help market analysts produce reports and analyse industry trends, or help business development managers find relevant companies for their product offering.

Start your project

Let's discuss a project like this for your company

Share a few details and our AI team will take it from there. Here is what happens next:

  1. We review your request and reply by email.
  2. A call with an AI expert to understand your problem, data and goals.
  3. A clear recommendation: the approach we suggest and a high-level estimate.

Prefer to pick a time yourself?

Schedule a call

By submitting, you agree to our privacy policy. We use your details only to reply to your request.

Discuss your project